Journal of Molecular Graphics and Modelling 20 (2001) 36–53
Molecular structure–property relationships for alkenes夽
Steven D. Nelson, Paul G. Seybold∗
Department of Chemistry, Wright State University, Dayton, OH 45435, USA
Received 28 November 2000; received in revised form 7 February 2001; accepted 7 February 2001
Abstract
Structure–property relationships were obtained for 11 physical and chemical properties (boiling points (bp), melting points (mp), molar
refractions (MR), molar volumes (MV), heats of combustion (HCKJ), molar heats of vaporization (HVMOL), flashpoints (FLASHK),
second virial coefficients (VIRC2), critical temperatures (Tc ), critical pressures (Pc ), and viscosities (VISC)) for a data set consisting
of 162 C4–C9 monoalkenes. Both molecular connectivity indices and ad hoc descriptors were tested as structural descriptors, and both
produced high-quality regression equations for most of the properties. As was observed in an earlier study of alkanes [J. Am. Chem. Soc.
110 (1988) 4186], mp were not well described by either descriptor set. For most properties, the mass/bulk of the molecule was found to
be the most important structural feature determining the property, suggesting that dispersion forces play a dominant role in determining
those properties influenced by intermolecular interactions. The amount of branching in the molecule and the nature of the double bond
environment were also found to be influential features. © 2001 Elsevier Science Inc. All rights reserved.
Keywords: Structure–property relationships; Alkenes; Principal component analysis (PCA)
1. Introduction
One of the most fundamental ideas of chemistry is that
the physical and chemical properties of a substance are determined, somehow, by its molecular structure — the term
‘structure’ taken here in its broadest sense to include both
geometric and electronic aspects. Historically, the difficulty
associated with this simple proposition has centered on how
to implement it in practice, since finding appropriate mathematical terms to describe ‘molecular structure’ has not been
easy. Such mathematical terms (‘descriptors’) are necessary
if quantitative relationships are to be constructed between
the structure of a compound and its properties. In principle, a full quantum mechanical treatment for a bulk property
such as the boiling point (bp) should yield a solution to this
conundrum, but in practice a full-blown, accurate quantum
mechanical treatment of a sizable collection of molecules is
presently out of the question. One turns, therefore, to simpler
approaches. One such approach is to attempt to identify suitable quantum chemical descriptors obtained from studies of
isolated molecules. This approach, although quite successful in some cases, has in many other cases produced only
mixed results. A simpler, alternative approach is to employ
夽 Taken in part from the Honors Thesis of SDN at Wright State University.
∗ Corresponding author. Tel.: +1-937-775-2407; fax: +1-937-775-2717.
E-mail address: [email protected] (P.G. Seybold).
topological descriptors, which fortunately often prove quite
adequate to the task and, moreover, highly informative.
Several benefits can be derived from structure–property
relationship studies. For example, unmeasured properties
of related compounds can be estimated using the equations
derived from a structure–property study. On a more fundamental level, a clearer understanding of the roles that
specific structural features play in determining properties
can be drawn from the equations. And once such an understanding is achieved, this information can be used to design
hypothetical structures that might have desirable property
values. Finally, the structure–property equations can serve
as a useful check on the accuracy of property values already
reported in the literature, some of which may be measured
incorrectly or misreported.
In an earlier report [1] we applied topological descriptors
in a study of eight physical properties of a set of normal and
branched alkanes. In that study, it was found that of the topological descriptors examined, molecular connectivity indices
[2,3] and ad hoc descriptors [1,4], were especially successful in yielding high-quality structure–property relationships.
Good regression equations were obtained for seven of the
physical properties of the alkanes (the melting points (mp),
traditionally a subtle and difficult property to represent, were
an exception). In this report, we employ these same descriptor types as structural measures for a study of the physical
and chemical properties of a set of monoalkenes, where a
new structural feature, the double bond, is introduced. Only
1093-3263/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved.
PII: S 1 0 9 3 - 3 2 6 3 ( 0 1 ) 0 0 0 9 9 - 7
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
a few previous QSPR studies have been devoted to the properties of this class of compounds [5–8], and these have generally been limited to single properties.
37
Molecular connectivity indices were originally developed
by Randic’ [15], Kier and Hall [2,3]. The form of these
descriptors is
m v
χt
2. Methods
The properties examined in this study were: bp, molar
refractions (MR), molar volumes (MV) at 20◦ C, heats of
combustion (HCKJ), mp, molar heats of vaporization (HVMOL) at 25◦ C, flashpoints (FLASHK), second virial coefficients (VIRC2) at 25◦ C, critical temperatures (Tc ), critical
pressures (Pc ), and viscosities (VISC) at 20◦ C. The property
values were taken from reference sources [9–14]. MV were
calculated as Mw /d, where Mw is the molecular weight, and
d is the density (g/cm3 ) at 20◦ C. MR were calculated using
the Lorentz–Lorenz expression
(n2 − 1)/(n20 + 2)
MR = 0
Mw /d
where n0 is the index of refraction.
The compounds examined and their shorthand abbreviations are listed in Appendix A. The property values for each
compound are shown in Appendix B. The parameter values
for each compound are shown in Appendix C.
Two types of parameters were used in this study: ad
hoc descriptors [1,4] and molecular connectivity indices
[2,3]. The ad hoc descriptors used were as follows: the
number of carbon atoms (NC) represents the mass or bulk
of the molecule. The square of the number of carbons
(NCSC) and the square root (NCSR) account for non-linear
aspects of the bulk dependence. The number of terminal
methyl groups (TM) is a measure of the amount of branching in the molecule. The number of paths of length three
carbon–carbon single bonds (P3S) is a steric index. The
descriptor exterior double bonds (DBE) indicates whether
the double bond is exterior or interior. The number of carbon atoms bonded to the double bond carbons (NCDB)
measures the amount of crowding in the double bond environment. Several additional ad hoc indices were examined,
but did not significantly improve the regression results.
where m is the order of the substructure, indicating the
number of bonds, and t is the substructure type indicator.
The substructure types included were path (p), cluster (c),
path/cluster (pc), double bond (DB), and total (t) [1]. Only
the valence (v) type indices, i.e. those which take explicit
account of heteroatoms and multiple bonds, have been used
in this study. The construction of these indices is described
in detail in the alkane study [1] and elsewhere [2,3]. The
molecular connectivity indices, used in this study, were calculated using Hall’s MOLCONN2 software package [16].
Here, these indices are identified as
XV(t)(m)(f )
where f is the functional form (I = inverse, SQ = square,
SR = square root). For example, XV1SR is the square root
of the first order valence index, and XVPC4 is the fourth order pc index. Also included were the numbers of three-bond
clusters (NXC3) and paths length of six bonds (NXP6). Altogether, 22 connectivity index terms were included in the
initial screening.
Regression equations and other statistical measures were
obtained using options in the SAS software package [17]
on the Wright State University IBM 8083E computer. The
final equations were selected on the basis of their standard
errors (S.E.) and F-statistics. Principal component analysis
(PCA) was used to determine the inherent dimensionality of
the groups of properties. Orthogonal and oblique rotations
did not significantly improve the results.
3. Results
3.1. Correlation analysis
The correlations among the properties examined are
shown in Table 1. As can be seen, most of the properties
Table 1
Correlations among the properties examined
Bp
MR
MV
HCKJ
HVMOL
FLASHK
VIRC2
Tc
Pc
VISC20
Mp
Bp
MR
MV
HCKJ
HVMOL
FLASHK
VIRC2
Tc
Pc
VISC20
Mp
1.000
0.967
0.946
0.970
0.993
0.933
−0.960
0.996
−0.950
0.659
0.664
1.000
0.992
0.996
0.921
0.905
−0.981
0.965
−0.940
0.824
0.640
1.000
0.992
0.903
0.844
−0.969
0.979
−0.958
0.804
0.614
1.000
0.938
0.897
−0.976
0.989
−0.958
0.823
0.525
1.000
0.878
−0.974
0.998
−0.920
0.602
0.482
1.000
−0.975
0.997
−0.825
0.431
0.518
1.000
−0.942
0.903
−0.981
−0.543
1.000
−0.954
0.925
0.607
1.000
−1.000
−0.467
1.000
0.844
1.000
38
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
Table 2
Multiple regression equation for the properties using ad hoc descriptors
Bp (◦ C)
−275.58(±3.51) + 139.31(±1.40) × NCSR + 7.49(±0.36) × NCDB − 7.64(±0.26) × TM − 5.93(±0.62) × DBE
+ 1.74(±0.19) × P3S; n = 162, r2 = 0.9941, S.E. = 2.28, F = 5252
MR (cm3 /mol)
20.65(±2.20) + 7.44(±0.35) × NC − 14.53(±1.76) × NCSR − 0.0862(±0.0130) × P3S; n = 156, r2 = 0.9976,
S.E. = 0.2224, F = 21114
MV (cm3 /mol)
26.27(±0.66) + 17.33(±0.12) × NC − 1.99(±0.11) × NCDB − 1.27(±0.07) × P3T + 0.87(±0.11) × TM; n = 156,
r2 = 0.9971, S.E. = 0.8547, F = 13105
HCKJ (kJ/mol)
84.57(±2.89) + 659.83(±0.45) × NC − 8.83(±0.53) × TM; n = 65, r2 = 1.0000, S.E. = 3.37, F = 1000000
Mp (◦ C)
−169.58(±7.06) + 0.9947(±0.1508) × NCSO; n = 57, r2 = 0.4416, S.E. = 17.43, F = 44
HVMOL (J/mol)
−27094(±1150) + 24224(±490) × NCSR + 1984(±192) × NCDB − 1972(±134) × TM − 1047(±237) × DBE;
n = 34, r2 = 0.9896, S.E. = 503, F = 690
FLASHK (K)
22.08(±27.94) + 92.78(±11.25) × NCSR; n = 18, r2 = 0.8096, S.E. = 5.81, F = 68
VIRC2 (cm3 /mol)
3630(±597) + 111.5(±6.3) × NCSO + 2427(±343) × NCSR − 68.5(±16.7) × NCDB; n = 14, r2 = 0.9978,
S.E. = 50.8, F = 1478
Tc (◦ C)
−63.73(±11.54) + 64.51(±3.90) × NC·2.30(±0.31) × NCSO − 10.42(±1.50) × DBE; n = 13, r2 = 0.9981,
S.E. = 2.45, F = 1607
Pc (MPa)
7.585(±0.319) − 1.789(±0.139) × NCSR; n = 12, r2 = 0.9432, S.E. = 0.128, F = 166
VISC20 (cP)
0.08661(±0.03813) + 0.00495(±0.00106) × NCSO + 0.00884(±0.00167) × TMSQ; n = 6, r2 = 0.9725,
S.E. = 0.01843, F = 53
are highly correlated with one another, with the exception
of mp, which is poorly correlated with the other properties.
VISC is another possible exception. The remaining nine
properties all have correlation coefficients greater than 0.82,
and the subset of bp, MR, MV, and HV all have correlations
greater than 0.90.
3.2. Multiple regression analysis
Table 2 shows the most successful ad hoc descriptor models with their coefficients of determination (r2 ), S.E., and
F-values, and Table 3 shows these measures for the most
successful molecular connectivity index models. Most of the
Table 3
Multiple regression equation for the properties using connectivity indices
Bp (◦ C)
15.04(±16.81) + 74.70(±6.08) × XV1 + 307.40(±15.99) × XVDB·81.79(±5.36) × XV0 + 28.49(±1.69) × XVP3
+ 51.20(±3.29) × XVZ + 157.89(±16.49) × XV1SR; n = 162, r2 = 0.9945, S.E. = 2.20, F = 46.80
MR (cm3 /mol)
1.91(±0.57) + 8.87(±0.14) × XV1 + 2.19(±0.04) × XV2 + 14.00(±1.26) × XVT;
n = 156, r2 = 0.9973, S.E. = 0.2374, F = 18527
MV (cm3 /mol)
7.79(±0.89) + 14.19(±0.36) × XV1 + 14.21(±0.24) × XV0 + 59.32(±1.58) × XVDB + 0.14(±0.26) × XVP3;
n = 156, r2 = 0.9973, S.E. = 0.8340, F = 13766
HCKJ (kJ/mol)
807.59(±10.34) + 1038.90(±3.81) × XV1 + 308.07(12.51) × XV2 + 103.80(±4.82) × XVP3 + 35.22(±2.68) × XVDB1;
n = 65, r2 = 0.9998, S.E. = 6.44, F = 90708
Mp (◦ C)
−414.50(±50.25) + 77.34(±12.22) × XV1 + 491.59(±107.89) × XVT; n = 57, r2 = 0.5510, S.E. = 15.77, F = 33
HVMOL (J/mol)
−19604(±11.63) + 27409(±852) × XV1SR + 2644(±246) × XVDB1 + 1950(±282) × NXP6; n = 34, r2 = 0.9884,
S.E. = 523, F = 852
FLASHK (K)
−113.57(±59.27) + 208.28(±30.67) × XV1SR + 21291(±64.76) × XVT; n = 18, r2 = 0.8982, S.E. = 44.5, F = 64
VIRC2 (cm3 /mol)
1310.1(±145.7) + 405.7(±11.5) × XV1SQ + 3421(±402) × XVT − 336(±46) × XVPC4; n = 14, r2 = 0.9982,
S.E. = 44.5, F = 1901
Tc (◦ C)
−440.18(±56.79) + 416.76(±35.72) × XV1SR + 308.33(±50.29) × XVT + 5.65(±1.22) × XVT12; n = 13, r2 = 0.9974,
S.E. = 2.90, F = 1151
Pc (MPa)
7.592(±0.336) + 1.051(±0.100) × XV0 + 0.141(±0.025) × XV1SQ + 4.135(±0.795) × XVDB2; n = 12, r2 = 0.9878,
S.E. = 0.066, F = 215
VISC20 (cP)
0.2466(±0.0053) + 0.0454(±0.0023) × NXC3 + 0.1034(±0.0108) × NXP6; n = 6, r2 = 0.9028, S.E. = 0.0094,
F = 208
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
Fig. 1. Plot of calculated (ad hoc) boiling points vs. experimental boiling points.
Fig. 2. Plot of calculated (molecular connectivity) boiling points vs. experimental boiling points.
39
40
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
Table 4
Results of the principal components analysis for the properties: eigenvalues with cumulative fractional variance reproduced
Factor
Three properties (156 compounds)
Four properties (57 compounds)
Nine properties (11 compounds)
1
2
3
4
5
2.937(0.979)
0.058(0.998)
0.005(1.000)
3.446(0.862)
0.501(0.987)
0.047(0.999)
0.005(1.000)
7.632(0.848)
1.079(0.968)
0.148(0.984)
0.075(0.993)
0.055(0.999)
properties are reasonably well modeled by these two types
of descriptors (Figs. 1 and 2), with the exception of the mp,
which were not well modeled by either descriptor set. The
FLASHK are less well modeled than most other properties,
but this is not entirely unexpected since this property is difficult to measure and some reported values may be relatively
inaccurate.
3.3. Factor analysis
Three factor analysis (PCA) studies were performed,
focusing on differing sets of properties. These studies
ranged from a ‘compound intensive’ study of just three
properties (bp, MR, and MV), for which a large number of experimental values (156) were available, to a
‘property intensive’ study including nine properties, for
which complete, matching data were available for only 11
compounds.
The results from the first factor analysis, using only bp,
MR and MV, are shown in Table 4. As can be seen above, the
Table 5
PCA factor loadings for the physical properties
Property
Factor
1
Three properties
Bp
MR
MV
0.981
0.997
0.990
2
3
0.191
−0.054
−0.135
0.014
−0.055
0.042
Four properties
Bp
MR
MV
Mp
0.976
0.982
0.972
0.766
−0.126
−0.174
−0.204
0.643
−0.177
0.056
0.109
0.015
Nine properties
Bp
MR
MV
Mp
HCKJ
VIRC2
Tc
Pc
VISC20
0.968
0.966
0.982
−0.062
0.998
−0.976
0.971
−0.954
0.996
0.163
−0.138
−0.093
0.993
−0.004
−0.143
0.106
0.078
−0.037
0.173
−0.110
−0.096
−0.100
0.017
−0.067
0.187
0.212
−0.036
three properties are highly correlated (Table 1) and a single
factor dominates, accounting for 97.9% of the variation in
these properties. As seen in Table 5, all three properties
load strongly on the first factor. The data set is quite large
(n = 156), and both low and high Mw compounds are well
represented.
Addition of mp to the analysis considerably reduced the
size of the data set, to 57 compounds, but the set still retained a good sampling of both high and low Mw compounds and remained large enough for reliable statistical
calculations. The results are shown in Tables 4 and 5. As
was seen in Table 1, the mp were not highly correlated with
the other properties. The property set now appears to be
represented by two factors (Table 4). The first factor accounts for 86.2% of the variance. Including the second factor, related to mp, brings this total to 98.7%. Reminiscent
of the alkane study [1], bp, MR, and MV load heavily on
the first factor, whereas mp loads strongly on the second
factor.
Addition of the heat of combustion, the VIRC2, and the
critical properties to the analysis markedly decreased the
common compound population, and the population became
strongly skewed toward the lower Mw monoalkenes. The
molar heat of vaporization was excluded to increase the number of compounds from 8 to 11. The results are shown in
Tables 4 and 5. This analysis is included only for the sake
Table 6
Modeling of factor scores using structural descriptors
Parameters
r2
S.E.
F
Factor 1
NC
NC, TMSQ
NC, TMSQ, DBE
XV1SR
XV1SR, XV0
XV1SR, XV0, NXP6
0.9947
0.9966
0.9986
0.9684
0.9936
0.9960
0.0726
0.0585
0.0382
0.1783
0.0803
0.0638
29223
22553
35347
4720
11929
12630
Factor 2
NCDB
NCDB, NCSR
NCDB, NCSR, P3S
XVTSQ
XVTSQ, XV1SR
XVTSQ, XV1SR, XV1
0.0757
0.1236
0.2225
0.1564
0.3121
0.3610
0.9701
0.9533
0.9063
0.9268
0.8446
0.8217
5
4
5
10
12
10
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
of completeness, since its statistical significance is questionable. However, it is interesting to note that the first factor
continued to dominate for most properties, accounting for
83.7% of the variance. Addition of the second factor brought
this up to 97.0%, and the addition of a third factor brought
this up to 99.8%.
In order to clarify, what the abstract factors represent, the
factor scores were modeled using the ad hoc and connectivity
descriptors. Factor 1 from the first analysis (n = 156) was
employed to obtain a model for this factor, and Factor 2
was taken from the second analysis (n = 57) for the same
purpose. The results are shown in Table 6. Factor 1 is clearly
a bulk factor, depending on the leading descriptors, whereas
Factor 2 was not well modeled by either of the descriptor
sets.
4. Discussion
The regression equations presented in Tables 2 and 3
are generally of high-quality for properties other than the
mp. Therefore, property values estimated on the basis of
these equations, with the exception of mp, should be sufficiently accurate for many practical purposes.
The ad hoc descriptor equations in Table 2 show the
relative influences of molecular mass/bulk (NC, NCSR
and NCSQ), branching (TM), steric factors (P3S), and the
double bond environment (DBE, NCDB) in determining
the properties studied. As can be seen from the Table, the
molecular mass/bulk clearly exerts the dominant influence
for properties other than mp, suggesting that dispersion
forces play a dominant role for those properties which depend on intermolecular forces. A similar conclusion was
reached in the earlier alkane study [1,4]. This is a reasonable
conclusion in the present case for bp, HVMOL, VIRC2, Tc ,
Pc , and VISC20. For MV the ‘mass/bulk’ dependence can
be attributed directly to the larger volume of compounds
with higher NC, as later modified by small corrections for
branching and steric influences. Likewise, MR depends
largely on the higher number of electrons in the larger
compounds. For the two strictly “chemical” properties, the
HCKJ and the flashpoints (FLASHK), the dependence on
the mass/bulk dimension is more accurately attributed to
the larger number of reacting bonds in the larger, higher NC
compounds.
Branching, steric factors, and the double bond environment exert smaller influences on the properties, as
demonstrated by the coefficients in the ad hoc regression
equations. Molecular branching, represented by the number of TM sequesters interior parts of these compounds
and reduces the extent of contact between neighboring
molecules. The latter effect is reflected in the positive influence of TM on the MV. Because dispersion forces are
strongly dependent on distance — the interaction energies
fall as 1/r6 , where r is the separation — a decrease in
the amount of close contact decreases the cohesive forces
41
experienced by the compounds. Therefore, bp and HVKJ
decrease as TM increases. Steric crowding, included here
by the parameter P3S, leads to a small reduction in the
MV, as reflected in the relatively small contributions to
MV and MR from P3S. Whether the double bond is exterior or interior and the NSDB influence the accessibility
of the double bond to its environment. When the double
bond is exterior the bp, for example, is reduced on average
by 6◦ C. Thus, DBE appear to exert a negative influence
on the intermolecular forces. The NCDB exerts a positive
influence on the intermolecular forces, possibly by effectively screening the effects of the double bond. Because
of the above influences, those properties that depend on
the strength of intermolecular forces, such as bp, HV, and
Tc , show positive dependences on NC, NCSQ, NCSR,
P3S, and NCDB, and negative dependences on TM and
DBE.
The failure of both parameter sets to model the mp is not
surprising, since this property was also not well modeled
by these same topological parameters in our earlier study
of the alkanes [1,4]. This illustrates the greater subtlety of
the melting transition as compared to the boiling and critical transitions. The latter transitions involve a direct dependence on the operative intermolecular forces, and so directly
reflect the strengths of these forces. The melting transition,
in contrast, maintains a condensed phase and involves a partial disruption of intermolecular orientations. Melting, thus,
depends on geometric and other factors that are not well
addressed by the present topological parameters. This dependence on shape and entropic factors, as opposed to a
simple intermolecular force dependence, is reflected in the
mp strong loading on the second factor, rather than the first
(mass/bulk related) factor, in the factor analysis. Dearden
has recently given a comprehensive review of mp predictions
[18].
Factor analysis for the monoalkenes shows that a single
factor dominates for most properties. In the worst case
examined (Analysis 3), Factor 1 still accounted for nearly
84% of the variance. Addition of a second factor in this
case raised the variance accounted to 97%. Factor 1 is
clearly related to the mass/bulk of the molecules. As can
be seen, in Table 6 both the ad hoc and connectivity index
mass/bulk dependent descriptors give a good account of
this factor. Factor 2, however, is a different story. From
the results in Table 5, it is obvious that Factor 2 is related to the mp, but this factor was not well modeled by
either the ad hoc or connectivity index descriptors employed in this study. Because of this, it is difficult to determine exactly what structural features are crucial for this
dimension.
Acknowledgements
We thank Stephen Peterangelo for checking some of the
regressions and preparing the figures.
42
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
Appendix A. Compounds
Observation no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
Name
Shorthand name
1-Butene
Cis-2-butene
Trans-2-butene
2-Methyl propene
1-Pentene
Cis-2-pentene
Trans-2-pentene
2-Methyl-1-butene
3-Methyl-1-butene
2-Methyl-2-butene
1-Hexene
Cis-2-hexene
Trans-2-hexene
Cis-3-hexene
Trans-3-hexene
2-Methyl-1-pentene
3-Methyl-1-pentene
4-Methyl-1-pentene
2-Methyl-2-pentene
3-Methyl-cis-2-pentene
3-Methyl-trans-2-pentene
4-Methyl-cis-2-pentene
4-Methyl-trans-2-pentene
2-Ethyl-1-butene
2,3-Dimethyl-1-butene
3,3-Dimethyl-1-butene
2,3-Dimethyl-2-butene
1-Heptene
Cis-2-heptene
Trans-2-heptene
Cis-3-heptene
Trans-3-heptene
2-Methyl-1-hexene
3-Methyl-1-hexene
4-Methyl-1-hexene
5-Methyl-1-hexene
2-Methyl-2-hexene
3-Methyl-cis-2-hexene
3-Methyl-trans-2-hexene
4-Methyl-cis-2-hexene
4-Methyl-trans-2-hexene
5-Methyl-cis-2-hexene
5-Methyl-trans-2-hexene
2-Methyl-cis-3-hexene
2-Methyl-trans-3-hexene
3-Methyl-cis-3-hexene
3-Methyl-trans-3-hexene
2-Ethyl-1-pentene
3-Ethyl-1-pentene
2,3-Dimethyl-1-pentene
2,4-Dimethyl-1-pentene
3,3-Dimethyl-1-pentene
3,4-Dimethyl-1-pentene
4,4-Dimethyl-1-pentene
1N4
2C4
2T4
2M1N3
1N5
2C5
2T5
2M1N4
3M1N4
2M2N4
1N6
2C6
2T6
3C6
3T6
2M1N5
3M1N5
4M1N5
2M2N5
3M2C5
3M2T5
4M2C5
4M2T5
2E1N4
23M1N4
33M1N4
23M2N4
1N7
2C7
2T7
3C7
3T7
2M1N6
3M1N6
4M1N6
5M1N6
2M2N6
3M2C6
3M2T6
4M2C6
4M2T6
5M2C6
5M2T6
2M3C6
2M3T6
3M3C6
3M3T6
2E1N5
3E1N5
23M1N5
24M1N5
33M1N5
34M1N5
44M1N5
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
Appendix A (Continued)
Observation no.
Name
Shorthand name
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
3-Ethyl-2-pentene
2,3-Dimethyl-2-pentene
2,4-Dimethyl-2-pentene
3,4-Dimethyl-cis-2-pentene
3,4-Dimethyl-trans-2-pentene
4,4-Dimethyl-cis-2-pentene
4, 4-Dimethyl-trans-2-pentene
2-Ethyl-3-methyl-1-butene
2,3,3-Trimethyl-1-butene
1-Octene
Cis-2-octene
Trans-2-octene
Cis-3-octene
Trans-3-octene
Cis-4-octene
Trans-4-octene
2-Methyl-1-heptene
3-Methyl-1-heptene
4-Methyl-1-heptene
5-Methyl-1-heptene
6-Methyl-1-heptene
2-Methyl-2-heptene
3-Methyl-cis-2-heptene
3-Methyl-trans-2-heptene
4-Methyl-cis-2-heptene
4-Methyl-trans-2-heptene
5-Methyl-cis-2-heptene
5-Methyl-trans-2-heptene
6-Methyl-cis-2-heptene
6-Methyl-trans-2-heptene
2-Methyl-cis-3-heptene
2-Methyl-trans-3-heptene
3-Methyl-cis-3-heptene
3-Methyl-trans-3-heptene
4-Methyl-cis-3-heptene
4-Methyl-trans-3-heptene
5-Methyl-cis-3-heptene
5-Methyl-trans-3-heptene
6-Methyl-cis-3-heptene
6-Methyl-trans-3-heptene
2-Ethyl-1-hexene
3-Ethyl-1-hexene
4-Ethyl-1-hexene
2,3-Dimethyl-1-hexene
2,4-Dimethyl-1-hexene
2,5-Dimethyl-1-hexene
3,3-Dimethyl-1-hexene
3,4-Dimethyl-1-hexene
3,5-Dimethyl-1-hexene
4,4-Dimethyl-1-hexene
4,5-Dimethyl-1-hexene
5,5-Dimethyl-1-hexene
3-Ethyl-cis-2-hexene
3-Ethyl-trans-2-hexene
3E2N5
23M2N5
24M2N5
34M2C5
34M2T5
44M2C5
44M2T5
2E3M1N4
233M1N4
1N8
2C8
2T8
3C8
3T8
4C8
4T8
2M1N7
3M1N7
4M1N7
5M1N7
6M1N7
2M2N7
3M2C7
3M2T7
4M2C7
4M2T7
5M2C7
5M2T7
6M2C7
6M2T7
2M3C7
2M3T7
3M3C7
3M3T7
4M3C7
4M3T7
5M3C7
5M3T7
6M3C7
6M3T7
2E1N6
3E1N6
4E1N6
23M1N6
24M1N6
25M1N6
33M1N6
34M1N6
35M1N6
44M1N6
45M1N6
55M1N6
3E2C6
3E2T6
43
44
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
Appendix A (Continued)
Observation no.
Name
Shorthand name
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
4-Ethyl-cis-2-hexene
4-Ethyl-trans-2-hexene
2,3-Dimethyl-2-hexene
2,4-Dimethyl-2-hexene
2,5-Dimethyl-2-hexene
3,4-Dimethyl-cis-2-hexene
3,4-Dimethyl-trans-2-hexene
3,5-Dimethyl-cis-2-hexene
3,5-Dimethyl-trans-2-hexene
4,4-Dimethyl-cis-2-hexene
4,4-Dimethyl-trans-2-hexene
4,5-Dimethyl-cis-2-hexene
4,5-Dimethyl-trans-2-hexene
5,5-Dimethyl-cis-2-hexene
5,5-Dimethyl-trans-2-hexene
3-Ethy1-3-hexene
2,2-Dimethyl-cis-3-hexene
2,2-Dimethyl-trans-3-hexene
2,3-Dimethyl-cis-3-hexene
2,3-Dimethyl-trans-3-hexene
2,4-Dimethyl-cis-3-hexene
2,4-Dimethyl-trans-3-hexene
2,5-Dimethyl-cis-3-hexene
2,5-Dimethyl-trans-3-hexene
3,4-Dimethyl-cis-3-hexene
3,4-Dimethyl-trans-3-hexene
2-n-Propyl-1-pentene
2-Isopropyl-1-pentene
2-Ethyl-3-methyl-1-pentene
2-Ethyl-4-methyl-1-pentene
3-Ethyl-2-methyl-1-pentene
3-Ethyl-3-methyl-1-pentene
3-Ethyl-4-methyl-1-pentene
2,3,3-Trimethyl-1-pentene
2,3,4-trimethyl-1-pentene
2,4,4-Trimethyl-1-pentene
3,3,4-Trimethyl-1-pentene
3,4,4-Trimethyl-1-pentene
3-Ethyl-2-methyl-2-pentene
3-Ethyl-4-methyl-cis-2-pentene
3-Ethyl-4-methyl-trans-2-pentene
2,3,4-Trimethyl-2-pentene
2,4,4-Trimethyl-2-pentene
3,4,4-Trimethyl-cis-2-pentene
3,4,4-Trimethyl-trans-2-pentene
2-Isopropyl-3-methyl-1-butene
2-Ethyl-3,3-dimethyl-1-butene
1-Nonene
Cis-2-nonene
Trans-2-nonene
Cis-3-nonene
Trans-3-nonene
Cis-4-nonene
Trans-4-nonene
4E2C6
4E2T6
23M2N6
24M2N6
25M2N6
34M2C6
34M2T6
35M2C6
35M2T6
44M2C6
44M2T6
45M2C6
45M2T6
55M2C6
55M2T6
3E3N6
22M3C6
22M3T6
23M3C6
23M3T6
24M3C6
24M3T6
25M3C6
25M3T6
34M3C6
34M3T6
2NP1N5
21P1N5
2E3M1N5
2E4M1N5
3E2M1N5
3E3M1N5
3E4M1N5
233M1N5
234M1N5
244M1N5
334M1N5
344M1N5
3E2M2N5
3E4M2C5
3E4M2T5
234M2N5
244M2N5
344M2C5
344M2T5
21P3M1N4
2E33M1N4
1N9
2C9
2T9
3C9
3T9
4C9
4T9
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
45
46
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
47
48
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
49
50
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
51
52
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
S.D. Nelson, P.G. Seybold / Journal of Molecular Graphics and Modelling 20 (2001) 36–53
References
[1] D.E. Needham, I.-C. Wei, P.G. Seybold, J. Am. Chem. Soc. 110
(1988) 4186–4194.
[2] L.B. Kier, L.H. Hall, Molecular Connectivity in Chemistry and Drug
Research, Academic Press, New York, 1976.
[3] L.B. Kier, L.H. Hall, Molecular Connectivity in Structure–Activity
Analysis, Wiley, New York, 1986.
[4] P.G. Seybold, M. May, U.A. Bagal, J. Chem. Educ. 64 (1987) 575–
581.
[5] R.H. Rohrbaugh, P.C. Jurs, Anal. Chem. 57 (1985) 2770–2773.
[6] P.J. Hansen, P.C. Jurs, Anal. Chem. 59 (1987) 2322–2327.
[7] S. Liu, R. Zhang, M. Liu, Z.J. Hu, Chem. Info. Comput. Sci. 37
(1997) 1146–1151.
[8] S.P. Verevkin, D. Wandschneider, A. Heintz, J. Chem. Eng. Data 45
(2000) 618–625.
[9] CRC Handbook of Chemistry and Physics, 68th Edition, CRC Press,
Boca Raton, 1987.
53
[10] I.B.D. Smith, R. Srivastava, Thermodynamic Data for Pure Compounds. Part A. Hydrocarbons and Ketones, Elsevier, New York, 1986.
[11] R.R. Dreisbach (Ed.), Physical Properties of Chemical Compounds.
Part II. Advances in Chemistry Series, Vol. 22, American Chemical
Society, Washington, DC, 1959.
[12] Selected Values of Properties of Hydrocarbons and Related
Compounds, A&M Research Foundation, Thermodynamic Research
Center, College Station, TX, 1985.
[13] M.P. Doss, Physical Constants of the Principal Hydrocarbons, 2nd
Edition, The Texas Company, New York, 1939.
[14] Aldrich Catalog of Fine Chemicals, Aldrich Chemical Co., New
York, 1981–1982.
[15] M. Randic’, J. Am. Chem. Soc. 97 (1975) 6609–6615.
[16] L.H. Hall, Hall Associates, Quincy, MA.
[17] SAS Institute, Box 8000, Cary, NC.
[18] J.C. Dearden, The prediction of melting point, in: M. Charton (Ed.),
Advances in Quantitative Structure Property Relationships, Vol. 2,
JAI Press, New York, 1999, pp. 127–175.